PSYC 394 Lecture Notes - Lecture 1: Partial Correlation, Standard Deviation, Linear Regression
Document Summary
Null hypothesis = h0 = an effect is absent. Alternative hypothesis = h1 = an effect is present. Psychologists are interested in results applicable to populations, but use samples due to limitations in data collection (e. g. inaccessibility to resources, subgroups). Sampling variation (aka. sampling error: random variability among samples taken from the same population, h0 : no difference between a and b. Computing significance: compute a test statistic (e. g. t, f, compare the value to sampling distributions. Central limit theorem (clt) implies regression towards the mean. Type 1 error ( ) : rejecting the h0 when it is true. Type 2 error ( ) : failing to reject the h0 when it is false. [a trade-off exists between errors during psychological research: decreasing the likelihood of a type 1 error increase the probability of a type 2 error, and vice versa. ] One-tailed test the hypothesis is directional ( a > b or a < b )
